crawler operating in local

database/common.py

@@ -0,0 +1,29 @@
+from sqlalchemy.ext.asyncio import create_async_engine, AsyncSession
+from sqlalchemy.orm import declarative_base, sessionmaker
+ENVIRONMENT = "dev"
+DATABASE_URL = ""
+
+# 비동기 엔진 생성
+engine = create_async_engine(
+    DATABASE_URL.replace("postgresql://", "postgresql+asyncpg://"),
+    echo=True if ENVIRONMENT == "dev" else False,
+    future=True
+)
+
+# 세션 팩토리 생성
+AsyncSessionLocal = sessionmaker(
+    engine, 
+    class_=AsyncSession, 
+    expire_on_commit=False
+)
+
+# Base 클래스 생성
+Base = declarative_base()
+
+# 데이터베이스 세션 의존성
+async def get_db():
+    async with AsyncSessionLocal() as session:
+        try:
+            yield session
+        finally:
+            await session.close()

database/models.py

@@ -0,0 +1,396 @@
+from sqlalchemy import Column, Integer, String, DateTime, Text, Boolean, ForeignKey, JSON, Float, BigInteger, Date
+from sqlalchemy.orm import relationship
+from sqlalchemy.sql import func
+from sqlalchemy.dialects.postgresql import UUID
+from app.database import Base
+import uuid
+
+class Company(Base):
+    """회사 모델"""
+    __tablename__ = "company"
+    
+    company_id = Column(Integer, primary_key=True, index=True)
+    company_name = Column(String, nullable=True)
+    
+    # 관계
+    users = relationship("User", back_populates="company")
+
+class User(Base):
+    """사용자 모델"""
+    __tablename__ = "users"
+    
+    id = Column(Integer, primary_key=True, index=True)
+    username = Column(String, unique=True, index=True, nullable=True)  # username 필드 추가
+    email = Column(String, unique=True, index=True, nullable=False)
+    hashed_password = Column(String, nullable=False)
+    is_active = Column(Boolean, default=True)
+    created_at = Column(DateTime(timezone=True), server_default=func.now())
+    updated_at = Column(DateTime(timezone=True), onupdate=func.now())
+    name = Column(String, nullable=False)  # name 필드 복구
+    refresh_token = Column(String, nullable=True)
+    company_id = Column(Integer, ForeignKey("company.company_id"), nullable=True)
+    address = Column(String, nullable=True)
+    description = Column(String, nullable=True)
+    role = Column(String, nullable=True)  # 사용자, 마케터, 프랜차이즈 대표, 관리자
+    
+    # 관계
+    company = relationship("Company", back_populates="users")
+    projects = relationship("Project", back_populates="owner")
+
+class AdStore(Base):
+    """광고 매장 모델"""
+    __tablename__ = "ad_store"
+    
+    id = Column(Integer, primary_key=True, index=True)
+    user_id = Column(Integer, ForeignKey("users.id"), nullable=True)
+    name = Column(String, nullable=True)
+    category = Column(String, nullable=True)
+    site_address = Column(String, nullable=True)
+    road_address = Column(String, nullable=True)
+    phone_number = Column(String, nullable=True)
+    naver_place_url = Column(String, nullable=True)
+    hashtags = Column(JSON, nullable=False, default=list)
+    description = Column(Text, nullable=True)
+    image_urls = Column(JSON, nullable=True)
+    created_at = Column(DateTime(timezone=True), server_default=func.now())
+    found_year = Column(Integer, nullable=True)
+    business_hours = Column(String, nullable=True)  # 매장 오픈, close 시간, 주말, 휴무 관련
+    employ_count = Column(Integer, nullable=True)  # 매장 직원수
+    specialties = Column(String, nullable=True)  # 사업 전문 분야
+    
+    # 관계
+    user = relationship("User", backref="ad_stores")
+    projects = relationship("Project", backref="ad_store")
+
+class Preset(Base):
+    """프리셋 모델"""
+    __tablename__ = "preset"
+    
+    id = Column(Integer, primary_key=True, index=True)
+    user_id = Column(Integer, ForeignKey("users.id"), nullable=False)
+    name = Column(String, nullable=True)
+    category = Column(String, nullable=True)
+    description = Column(String, nullable=True)
+    favorite = Column(Boolean, default=False)
+    hashtag = Column(JSON, nullable=True)
+    image_urls = Column(JSON, nullable=True)
+    created_at = Column(DateTime(timezone=True), server_default=func.now())
+    last_used_at = Column(DateTime(timezone=True), nullable=True)
+    title = Column(String, nullable=True)
+    
+    # 관계
+    user = relationship("User", backref="presets")
+    projects = relationship("Project", back_populates="preset")
+
+class Template(Base):
+    """템플릿 모델"""
+    __tablename__ = "template"
+    
+    id = Column(Integer, primary_key=True, index=True)
+    user_id = Column(Integer, ForeignKey("users.id"), nullable=True)
+    name = Column(String, nullable=True)
+    description = Column(String, nullable=True)
+    thumbnail = Column(String, nullable=True)
+    template_url = Column(String, nullable=True)
+    category = Column(String, nullable=True)
+    rating = Column(Float, nullable=True)  # real type in PostgreSQL maps to Float
+    favortie = Column(Boolean, default=False)  # Note: typo in DB column name "favortie"
+    hashtags = Column(JSON, nullable=True)
+    created_at = Column(DateTime(timezone=True), server_default=func.now())
+    usage_count = Column(Integer, nullable=True)
+    
+    # 관계
+    user = relationship("User", backref="templates")
+    projects = relationship("Project", back_populates="template")
+
+class Project(Base):
+    """프로젝트 모델"""
+    __tablename__ = "projects"
+    
+    id = Column(Integer, primary_key=True, index=True)
+    title = Column(String, nullable=False)
+    description = Column(Text)
+    target_url = Column(String, nullable=False)  # 크롤링 대상 URL
+    status = Column(String, default="created")  # created, crawling, processing, completed, failed
+    owner_id = Column(Integer, ForeignKey("users.id"))
+    created_at = Column(DateTime(timezone=True), server_default=func.now())
+    
+    # Template and Store references (from database)
+    template_id = Column(Integer, ForeignKey("template.id"), nullable=True)
+    store_id = Column(Integer, ForeignKey("ad_store.id"), nullable=True)
+    preset_id = Column(Integer, ForeignKey("preset.id"), nullable=True)
+    
+    # YouTube 관련 필드
+    youtube_video_id = Column(String(50), nullable=True)
+    youtube_url = Column(String(255), nullable=True)
+    youtube_upload_status = Column(String(50), nullable=True)
+    youtube_privacy_status = Column(String(20), nullable=True)
+    updated_at = Column(DateTime(timezone=True), onupdate=func.now())
+    
+    # 관계
+    owner = relationship("User", back_populates="projects")
+    template = relationship("Template", back_populates="projects")
+    preset = relationship("Preset", back_populates="projects")
+    crawl_data = relationship("CrawlData", back_populates="project", uselist=False)
+    video = relationship("Video", back_populates="project", uselist=False)
+    generated_content = relationship("GeneratedContent", back_populates="project", uselist=False)
+    task_statuses = relationship("TaskStatus", back_populates="project")
+
+class CrawlData(Base):
+    """크롤링 데이터 모델"""
+    __tablename__ = "crawl_data"
+    
+    id = Column(Integer, primary_key=True, index=True)
+    project_id = Column(Integer, ForeignKey("projects.id"))
+    
+    # Naver Map 데이터
+    business_name = Column(String)
+    address = Column(String)
+    phone = Column(String)
+    business_type = Column(String)
+    coordinates = Column(JSON)  # {"lat": float, "lng": float}
+    
+    # 추가 장소 정보 (crawler4)
+    business_hours = Column(String)  # 영업시간
+    homepage = Column(String)  # 홈페이지 URL
+    place_description = Column(Text)  # 장소 설명
+    
+    # 블로그 데이터
+    blog_urls = Column(JSON)  # 수집된 블로그 URL 리스트
+    blog_content = Column(Text)  # 추출된 블로그 내용
+    advertising_keywords = Column(JSON)  # 광고 키워드 리스트
+    
+    # 상세 블로그 정보 (crawler4)
+    blog_details = Column(JSON)  # detailed_blogs 전체 저장
+    map_blog_reviews = Column(JSON)  # 지도 블로그 리뷰
+    
+    # 이미지 데이터
+    image_urls = Column(JSON)  # 수집된 이미지 URL 리스트
+    stored_images = Column(JSON)  # 저장된 이미지 파일 경로 리스트
+    
+    # 크롤링 메타데이터
+    crawl_metadata = Column(JSON)  # 통계 정보, 수집 시간 등
+    
+    created_at = Column(DateTime(timezone=True), server_default=func.now())
+    
+    # 관계
+    project = relationship("Project", back_populates="crawl_data")
+
+class GeneratedContent(Base):
+    """생성된 콘텐츠 모델"""
+    __tablename__ = "generated_content"
+    
+    id = Column(Integer, primary_key=True, index=True)
+    project_id = Column(Integer, ForeignKey("projects.id"))
+    
+    # 가사 데이터
+    lyrics = Column(Text)
+    lyrics_prompt = Column(Text)  # 사용된 프롬프트
+    lyrics_generation_time = Column(Integer, default=0)  # 가사 생성 시간(초)
+    lyrics_token_count = Column(Integer, default=0)  # 사용된 토큰 수
+    lyrics_model_used = Column(String, default="gpt-3.5-turbo")  # 사용된 모델
+    
+    # BGM 데이터
+    bgm_url = Column(String)
+    bgm_file_path = Column(String)  # 저장된 BGM 파일 경로
+    bgm_duration = Column(Float, default=30.0)  # BGM 길이(초)
+    bgm_generation_time = Column(Integer, default=0)  # BGM 생성 시간(초)
+    bgm_model_used = Column(String, default="mureka-v6")  # 사용된 모델
+    
+    created_at = Column(DateTime(timezone=True), server_default=func.now())
+    
+    # 관계
+    project = relationship("Project", back_populates="generated_content")
+
+class Video(Base):
+    """동영상 모델"""
+    __tablename__ = "videos"
+    
+    id = Column(Integer, primary_key=True, index=True)
+    project_id = Column(Integer, ForeignKey("projects.id"))
+    
+    # 동영상 파일 정보
+    file_path = Column(String, nullable=False)
+    file_size = Column(Integer)  # 바이트 단위
+    duration = Column(Integer)  # 초 단위
+    resolution = Column(String)  # "1920x1080" 형식
+    
+    # 생성 설정
+    template_type = Column(String, default="default")
+    generation_settings = Column(JSON)  # 동영상 생성 설정
+    
+    # 상태
+    status = Column(String, default="processing")  # processing, completed, failed
+    error_message = Column(Text)
+    
+    created_at = Column(DateTime(timezone=True), server_default=func.now())
+    completed_at = Column(DateTime(timezone=True))
+    
+    # 관계
+    project = relationship("Project", back_populates="video")
+
+class TaskStatus(Base):
+    """작업 상태 추적 모델"""
+    __tablename__ = "task_status"
+
+    id = Column(Integer, primary_key=True, index=True)
+    task_id = Column(String, unique=True, index=True, nullable=True)  # Celery 태스크 ID
+    project_id = Column(Integer, ForeignKey("projects.id"), nullable=True)
+    task_type = Column(String, nullable=True)  # crawling, content_generation, video_generation
+    status = Column(String, nullable=True)  # pending, started, progress, success, failure
+    progress = Column(Integer, nullable=True)  # 0-100
+    current_step = Column(String, nullable=True)
+    result = Column(JSON, nullable=True)
+    error_message = Column(Text, nullable=True)
+
+    created_at = Column(DateTime(timezone=True), server_default=func.now())
+    updated_at = Column(DateTime(timezone=True), nullable=True)
+
+    # 관계
+    project = relationship("Project", back_populates="task_statuses")
+
+class UserActivityLog(Base):
+    """사용자 활동 로그 모델"""
+    __tablename__ = "user_activity_logs"
+    
+    id = Column(Integer, primary_key=True, index=True)
+    user_id = Column(Integer, ForeignKey("users.id"))
+    action = Column(String, nullable=False)  # login, logout, create_project, etc.
+    details = Column(JSON)  # 추가 정보
+    ip_address = Column(String)
+    user_agent = Column(String)
+    
+    created_at = Column(DateTime(timezone=True), server_default=func.now())
+    
+    # 관계
+    user = relationship("User")
+
+class SystemStatistics(Base):
+    """시스템 통계 모델"""
+    __tablename__ = "system_statistics"
+    
+    id = Column(Integer, primary_key=True, index=True)
+    stat_date = Column(Date, unique=True, nullable=False)
+    
+    # 사용자 통계
+    total_users = Column(Integer, nullable=True)
+    active_users = Column(Integer, nullable=True)
+    new_users = Column(Integer, nullable=True)
+    
+    # 프로젝트 통계
+    total_projects = Column(Integer, nullable=True)
+    completed_projects = Column(Integer, nullable=True)
+    failed_projects = Column(Integer, nullable=True)
+    
+    # 동영상 통계
+    total_videos = Column(Integer, nullable=True)
+    total_video_size = Column(BigInteger, nullable=True)  # bytes
+    new_videos = Column(Integer, nullable=True)  # 이번달 새로 생성한 비디오 갯수
+    
+    # 템플릿 통계
+    total_templates = Column(Integer, nullable=True)  # 전체 템플릿 갯수
+    
+    # 성능 통계
+    avg_crawling_time = Column(Float, nullable=True)  # seconds
+    avg_content_generation_time = Column(Float, nullable=True)  # seconds
+    avg_video_rendering_time = Column(Float, nullable=True)  # seconds
+    
+    # 증가율 통계 (월단위)
+    user_growth_rate = Column(Float, nullable=True)  # 사용자 증가량-월단위
+    template_growth_rate = Column(Float, nullable=True)  # 템플릿 생성 증가율 - 월단위
+    video_growth_rate = Column(Float, nullable=True)  # 비디오 생성 증가율 -월단위
+    media_growth_rate = Column(Float, nullable=True)  # 미디어 에셋 증가율 - 월단위
+    
+    created_at = Column(DateTime(timezone=True), server_default=func.now())
+    updated_at = Column(DateTime(timezone=True), onupdate=func.now())
+
+
+class SystemUsage(Base):
+    """시스템 사용량 모델"""
+    __tablename__ = "system_usage"
+    
+    id = Column(Integer, primary_key=True, index=True)
+    stat_date = Column(Date, unique=True, nullable=False)
+    
+    # 스토리지 사용량 (Giga 단위)
+    image_usage = Column(Float, nullable=True)  # 이미지 스토리지 사용량 - Giga 단위
+    video_usage = Column(Float, nullable=True)  # 비디오 스토리지 사용량 - Giga 단위
+    music_usage = Column(Float, nullable=True)  # 노래 스토리지 사용량 - Giga 단위
+    extra_usage = Column(Float, nullable=True)  # 기타 스토리지 사용량 - Giga 단위
+    
+    created_at = Column(DateTime(timezone=True), server_default=func.now())
+
+
+class SystemAvailability(Base):
+    """시스템 가용성 모델"""
+    __tablename__ = "system_availability"
+    
+    id = Column(Integer, primary_key=True, index=True)
+    stat_date = Column(Date, unique=True, nullable=False)
+    
+    # 가용성 지표 (퍼센트 단위 - smallint로 0-100 범위)
+    ai_availability = Column(Integer, nullable=True)  # AI 엔진 가동율
+    dbase_availability = Column(Integer, nullable=True)  # 데이터베이스 가동율
+    storage_availability = Column(Integer, nullable=True)  # 스토리지 서버 가동율
+    renderer_availability = Column(Integer, nullable=True)  # 비디오 Renderer 가동율
+    
+    created_at = Column(DateTime(timezone=True), server_default=func.now())
+
+
+class MediaAsset(Base):
+    """미디어 자산 모델"""
+    __tablename__ = "media_assets"
+    __mapper_args__ = {"confirm_deleted_rows": False}
+    
+    id = Column(Integer, primary_key=True, index=True)
+    name = Column(String, nullable=False)
+    type = Column(String, nullable=False)
+    size = Column(String)
+    file_size_bytes = Column(Integer)
+    dimensions = Column(String)
+    width = Column(Integer)
+    height = Column(Integer)
+    duration = Column(String)
+    duration_seconds = Column(Float)
+    thumbnail = Column(String)
+    file_path = Column(String)
+    url = Column(String)
+    mime_type = Column(String)
+    tags = Column(JSON)
+    usage_count = Column(Integer, default=0)
+    owner_id = Column(Integer, ForeignKey("users.id"))
+    meta_info = Column(JSON)  # metadata는 SQLAlchemy 예약어이므로 변경
+    created_at = Column(DateTime(timezone=True), server_default=func.now())
+    updated_at = Column(DateTime(timezone=True), onupdate=func.now())
+    
+    # 관계
+    owner = relationship("User", backref="media_assets")
+
+class Asset(Base):
+    """미디어 자산 모델"""
+    __tablename__ = "media_assets"
+    __mapper_args__ = {"confirm_deleted_rows": False}
+    
+    id = Column(Integer, primary_key=True, index=True)
+    name = Column(String, nullable=False)
+    type = Column(String, nullable=False)
+    size = Column(String)
+    file_size_bytes = Column(Integer)
+    dimensions = Column(String)
+    width = Column(Integer)
+    height = Column(Integer)
+    duration = Column(String)
+    duration_seconds = Column(Float)
+    thumbnail = Column(String)
+    file_path = Column(String)
+    url = Column(String)
+    mime_type = Column(String)
+    tags = Column(JSON)
+    usage_count = Column(Integer, default=0)
+    owner_id = Column(Integer, ForeignKey("users.id"))
+    meta_info = Column(JSON)  # metadata는 SQLAlchemy 예약어이므로 변경
+    created_at = Column(DateTime(timezone=True), server_default=func.now())
+    updated_at = Column(DateTime(timezone=True), onupdate=func.now())
+    
+    # 관계
+    owner = relationship("User", backref="media_assets")

main.py

@@ -34,10 +34,7 @@ async def main():
     print(f"  - 허용 포맷: {', '.join(sorted(image_filter.allowed_formats))}")
     print(f"  - 크기 조건: {'AND' if image_filter.require_both_dimensions else 'OR'}")
     
-    # 프로젝트 디렉토리 설정
-    if platform.system() == 'Windows':
-        project_dir = config.get('paths.project_dir_windows', r'C:\CrawlingData')
-    else:
+    # 프로젝트 디렉토리 설정)
     project_dir = os.path.join(os.path.dirname(os.path.realpath(__file__)), 'data')
     
     # 크롤러 초기화

requirements.txt

@@ -4,5 +4,8 @@ selenium
 requests
 pillow
 azure-identity
+azure-keyvault-secrets
+azure-storage-blob
+azure-storage-queue
 azure-mgmt-resource
 azure-mgmt-compute

selenium_crawler/naver_integrated_crawler.py

@@ -23,7 +23,7 @@ from .naver_blog_crawler import NaverBlogCrawler, ImageFilterConfig
 
 
 class NaverIntegratedCrawler:
-    def __init__(self, project_dir: str = r'/home/jhyeu/workspace/data', 
+    def __init__(self, project_dir:str, 
                  image_filter: ImageFilterConfig = None):
         """
         초기화

utils/azure.py

@@ -0,0 +1,87 @@
+import requests
+import json
+
+from azure.storage.blob import BlobServiceClient, ContentSettings
+from azure.storage.queue import QueueClient
+from azure.identity import DefaultAzureCredential
+from azure.keyvault.secrets import SecretClient
+from azure.identity import DefaultAzureCredential
+
+KEY_VAULT_NAME = "ado2-keys"
+def get_keyvault_client():
+    key_vault_uri = f"https://{KEY_VAULT_NAME}.vault.azure.net/"
+    
+    # DefaultAzureCredential이 Managed Identity를 자동으로 사용
+    credential = DefaultAzureCredential()
+    return SecretClient(vault_url=key_vault_uri, credential=credential)
+
+# Secret Key
+BLOB_ACCOUNT_URL_KEY = "ado2-blob-account-url"# https://mystorageaccount.blob.core.windows.net
+QUEUE_URL_KEY = "ado2-queue-url"
+
+# Plain text
+BLOB_CONTAINER_NAME = "ado2-media-public-access"
+BLOB_MEDIA_FOLDER = "ado2-media-original"
+
+def az_storage_upload_ado2_media(data, remote_file_path:str):
+    secret_client = get_keyvault_client()
+    account_url = secret_client.get_secret(BLOB_ACCOUNT_URL_KEY).value
+    # Managed Identity 사용
+    credential = DefaultAzureCredential()
+    blob_service_client = BlobServiceClient(account_url=account_url, credential=credential)
+
+    blob_client = blob_service_client.get_blob_client(container=BLOB_CONTAINER_NAME, blob=f"{BLOB_MEDIA_FOLDER}/{remote_file_path}")
+    
+    # with open(local_file_path, 'rb') as data:
+    #     blob_client.upload_blob(data, overwrite=True)
+    blob_client.upload_blob(data, overwrite=True)
+    return blob_client.url
+
+
+def az_storage_read_ado2_media(remote_file_path:str):
+    secret_client = get_keyvault_client()
+    account_url = secret_client.get_secret(BLOB_ACCOUNT_URL_KEY).value
+    # Managed Identity 사용
+    credential = DefaultAzureCredential()
+    blob_service_client = BlobServiceClient(account_url=account_url, credential=credential)
+
+    blob_client = blob_service_client.get_blob_client(container=BLOB_CONTAINER_NAME, blob=f"{BLOB_MEDIA_FOLDER}/{remote_file_path}")
+    
+    blob_data = blob_client.download_blob()
+    return blob_data
+
+def az_storage_get_ado2_media_list(remote_folder_path:str) -> list[str]:
+    secret_client = get_keyvault_client()
+    account_url = secret_client.get_secret(BLOB_ACCOUNT_URL_KEY).value
+    # Managed Identity 사용
+    credential = DefaultAzureCredential()
+    blob_service_client = BlobServiceClient(account_url=account_url, credential=credential)
+
+    prefix = f"{BLOB_MEDIA_FOLDER}/{remote_folder_path}"
+    blob_media_list = blob_service_client.get_container_client(BLOB_CONTAINER_NAME).list_blobs(name_starts_with=prefix)
+    return blob_media_list
+
+def az_add_queue_message(body:dict):
+    secret_client = get_keyvault_client()
+    queue_url = secret_client.get_secret(QUEUE_URL_KEY).value
+    credential = DefaultAzureCredential()
+    queue_client = QueueClient.from_queue_url(queue_url, credential=credential)    
+    queue_client.send_message(json.dumps(body, ensure_ascii=False))
+
+def az_get_queue_message():
+    secret_client = get_keyvault_client()
+    queue_url = secret_client.get_secret(QUEUE_URL_KEY).value
+    credential = DefaultAzureCredential()
+    queue_client = QueueClient.from_queue_url(queue_url, credential=credential)
+
+    # 메시지 읽기
+    messages = queue_client.receive_messages()
+    data_list = []
+    for message in messages:
+        # 메시지 처리
+        data = json.loads(message.content)
+        print(f"받은 메시지: {data}")
+        # 처리 후 삭제
+        queue_client.delete_message(message.id, message.pop_receipt)
+        data_list.append(data)
+    return data_list